Examples

The example below although written for Linux, should work on another operating system if adjusted accordingly.

Tuning DE and PSO to the CEC‘05 problems under multiple OFE budgets

For this example, a virtual machine running Ubuntu 14.04 will be setup to tune DE and PSO to the CEC‘05 problems using tMOPSO.

On your new virtual machine, with a fresh Ubuntu 14.04 installed, in a terminal

$ sudo apt-get install python-numpy python-scipy python-matplotlib   # python modules which optTune depends on
$ sudo apt-get install python-setuptools   #install optTune for the local user
$ easy_install --user optTune

f2py and batchOpenMPI are used for speed: f2py as to easily interface with the DE and PSO fortran codes, and batchOpenMPI for parallelization. For these you will need the following software

$ sudo apt-get install gfortran openmpi-bin openmpi-common libopenmpi-dev
$ easy_install --user mpi4py
$ easy_install --user batchOpenMPI

Next download and extract the CEC 2005 tuning example files located at cec05examples.tar.gz

$ wget https://pythonhosted.org/optTune/cec05examples.tar.gz
$ tar -xf cec05examples.tar.gz
$ cd cec05examples/
$ ls

The first thing to do is to compile and then validate the fortran codes

$ ./compile.sh
$ sudo apt-get install octave #octave is used in the testing scripts ...
$ ./test_cec2005problems.py

Examine the testing data, and if acceptable

# to tune DE using 4 processors
$ mpirun -np 4 tune_DE_via_tMOPSO.py
# or to tune PSO
$ mpirun -np 4 tune_PSO_via_tMOPSO.py

Have fun.